Inventário de potencial de abuso infantil - CAP: adaptação transcultural, fidedignidade e validade para o Brasil

Abstract

This study aimed at developing a cross-validation of the Child Abuse Potential Inventory (CAP) to Brazil. The specific objectives were to verify the semantic equivalence of the measure, to investigate and to assess the content validity, and to investigate and assess the construct validity using Item Factor Analysis and reliability. To evaluate the general and referential meaning of the CAP Inventory s items, 6 researchers translated the CAP Inventory and assessed it using the Assessing Semantic Equivalence Form. An English teacher also conducted a back-translation, and, finally, 13 parents evaluated the comprehension of each item. In terms of the general meaning, 89.4% of the items were considered without the need of changes. The Portuguese version of the CAP Inventory showed 81.4% of agreement between judges about the referential meaning of items. In general, 19 items (12%) were altered. To evaluate the content validity, three expert researchers participated as judges. By using the Assessing Content Validity Form, it was verified that when there was only one concept in an item, the judges agreed above 80% in 9 (31.1%) items. When there were two concepts, the judges agreed above 80% in 4 items (10.8%), and when there were 3 concepts, the judges did not agree above 80% in none of the items. In terms of construct validity, 135 parents were selected using a random sample procedure, and they answered the final version of the CAP Inventory, and a demographic questionnaire. Data from other Brazilian studies with the same version of the CAP were used to expand the analysis. In terms of participants profile for the present data: mean parent age of 35.9 years, female, married, less of four years of education, 2.5 children (mean). Demographic data were similar to other Brazilian studies and the North-American sample, except for years of education which was higher for the North-American sample. To analyze data from the CAP Inventory, a cross-adaptation of the validity scales was needed to define the cutscore for Brazil. The mean score for potential child abuse in the Brazilian sample was 180.1(102. 5). This result was higher than the mean score in the North-American sample which was 91 (75). Using the Principal-Components Factor Analysis, a structure of 5 factors was chosen: distress, unhappiness, rigidity, problems with child and others, and discipline. The resulting structure showed similarities to the original factor analysis of the English CAP Inventory, except for the discipline factor. It was hypothesized that the Brazilian version of the CAP Inventory could present an alternative factor structure to the original North-American version. Other studies should be developed to verify this hypothesis. The Crombach alpha of internal consistency was 0.95. In general, data showed positive results of semantic equivalence and content validity, construct validity and reliability, what confirms the hypothesis of viability of cross-adaptation of the CAP Inventory to Brazil. However, other studies should be conducted to create a strong body of knowledge that it will allow analysis of tendencies and relationships between results.